SamuKata
Voxta
Voxta

patreon


Some tips for Voxta Cloud

We made a tutorial video to help you get setup! Remember to also checkout doc.voxta.ai for other tutorials and documentation.

Also, some useful things to keep in mind:

We're happy to see such positive feedback, and the system is holding up! We're very excited to see how local services and privacy-first cloud services can coexist in the future, so we can democratize AI characters for people with older GPUs or who want to run 3D avatars, while serving the need to host your own services. What a crazy time!

Some tips for Voxta Cloud

Comments

Sorry but how do I check how many credits I have left?

Qatar Iron

Hey! https://portal.voxta.ai/

Voxta

Hi there. New to your Patreon and im trying to set up Voxta on my PC. You guys Mention to download the latest release for the Voxta server?? Could you please just link it here for me? Ty

John

You should have received a message when signing up. There is documentation and video tutorials at doc.voxta.ai :)

Voxta

how do i use voxta ?

Bauti

you should come on the discord if you need help troubleshooting if you didn't try that. voxta runs great!

Charlie Brickner

Honestly, too much hassle to make this work. When it worked before it wasn't stable, voxta VaM scenes would stop responding to voice , voice response would became repetitive after while.. I cant even use latest version. I gave up, spent too much time trying to make it work.

BopDobop

Download is?

Nathan Muntner

General question about future development. Will the future 'VisionCapture' services have the ability to feed images from both vam and webcams into voxta with clip/segment magic in a1111 or such? BTW Macaroni Maid 7b is freaking awesome. It's quick and fauxmaginative. And how the HELL did I not realize you simply need to drop a 6+ second audio clip into a directory to have coqui clone it?? Black goddamned magic I say. Elevenlab has some nice voices, and coqui is happy to work with their samples. That's really nice for lots of HQ local options. BTW is there a way to automatically delete generated history from Elevenlabs? It takes forever to click through and delete a long conversation.

Smumblepooch

For custom models, you can also try running the benchmarks (more, advanced, benchmarks) see how it fares. And if you don't use the app itself, you can save some processing by turning off action inference (used for avatars and for the smileys while in a conversation)!

Acid Bubbles

Thanks! I just realized that page actually has what I'm looking for, and I'll give that one a go. I've had some great interactions so far with TheBloke_CAMEL-13B-Role-Playing-Data-AWQ, and I've had a couple of interactions where it was consistent, imaginative, agreeable, and enthusiastic. I need to spend some time fleshing out my memory books. I was able to connect the VAM plug-in and have a two-way voice conversation with a character with the lip text syncing on, and it was goddamned impressive. Today I used the desktop app to have a background conversation while working, and even without an optimized programming model it was really informative on some subjects while still having a.. fauxonality. :)

Smumblepooch

Hey this is Reign from the Voxta team. A personal recommendation for a good LLM is Macaroni Maid 7b. The model is super fast and very smart. Action inference works well and from my tests this is the best model we've got. It's my "daily driver". Give it a shot.

Voxta

We try to keep a list here: https://doc.voxta.ai/docs/large-language-models/ and you can validate whether your favorite LLM model works using the Voxta Benchmarks! There's a lot of moving pieces, bolting OpenAI to a TTS is probably not very hard, and it's fun to build! What we're trying to do here is bind that to an SDK to connect to games, home automation systems, all with a companion / assistance that's unrestricted. I hope your friend will come say hi, too :)

Voxta

You should be all set now! If anyone else has trouble, please click on the Sync Patreon Benefits button in the portal, or reach out to us!

Voxta

You should be all set now! If anyone else has trouble, please click on the Sync Patreon Benefits button in the portal, or reach out to us!

Voxta

Okay, I've got an LLM running on an 3090, with voxta & VAM on a 4090 connected to it. Elevenlabs, etc. all set up and working. And.. damn. Impressive. The docs link REALLY helped tweak LLM settings. One suggestion might be a list of good LLM's (e.g. THEBLOKE_*) and a shortlist of their settings. While I feel much more informed about what the settings do after reading the docs, it's still nice to have a cheat sheet. :) It needs to be.. a lot of things for mass consumption. But the pieces are in place. My friend was posturing about the ability to eventually do this the other day. Next time he's over I'm going to let it surprise him by saying hi. :D

Smumblepooch

well I got nothing

AlvinNeil

You have the Companion tier! Benefits: Voxta Server, 5000 Voxta Cloud Credits (Companion Tier), Voxta Virt-A-Mate Plugin, Voxta Virt-A-Mate Scenes

AlvinNeil


More Creators